A 3D Real-Scale, Biomimetic, and Biohybrid Model of the Blood-Brain Barrier Fabricated through Two-Photon Lithography

Abstract

The investigation of the crossing of exogenous substances through the blood-brain barrier (BBB) is object of intensive research in biomedicine, and one of the main obstacles for reliable in vitro evaluations is represented by the difficulties at the base of developing realistic models of the barrier, which could resemble as most accurately as possible the in vivo environment. Here, for the first time, a 1:1 scale, biomimetic, and biohybrid BBB model is proposed. Microtubes inspired to the brain capillaries were fabricated through two-photon lithography and used as scaffolds for the co-culturing of endothelial-like bEnd.3 and U87 glioblastoma cells. The constructs show the maturation of tight junctions, good performances in terms of hindering dextran diffusion through the barrier, and a satisfactory trans-endothelial electrical resistance. Moreover, a mathematical model is developed, which assists in both the design of the 3D microfluidic chip and its characterization. Overall, these results show the effective formation of a bioinspired cellular barrier based on microtubes reproducing brain microcapillaries to scale. This system will be exploited as a realistic in vitro model for the investigation of BBB crossing of nanomaterials and drugs, envisaging therapeutic and diagnostic applications for several brain pathologies, including brain cancer.

Keywords: biohybrid systems; biomimetics; blood-brain barrier; microfluidics mathematical models; two-photon lithography.

Figure 1

Bioinspired microfluidic system design and microfabrication. a) Schematic of the porous tube (mimicking a microcapillary) that simultaneously scaffolds the cells and allows for species transport towards the external environment. b) Schematic (top view) of the microfluidic chip devised to operate multiple porous tubes in parallel. c) Scheme depicting the two-photon lithography fabrication of the microfluidic system with porous tubular-shaped microcapillaries. d) Design of the microfluidic system consisting of the microcapillaries arranged in parallel and connected to the inflow and the outflow reservoirs. e) High magnification of the designed porous microcapillaries. Representative scanning electron microscopy images f) of the obtained microfluidic chip, g) of the tubes, and h) of the pores, highlighting the high reproducibility and resolution of the fabrication technique.

Figure 2

Modeling of the microfluidic system. a) Intensity plot of the fluid speed numerically computed on a simplified 2D geometry (half domain, by symmetry). b) Detail showing the struts supporting the junctions, and the considered cut-sections (δ = 0.01 μm). c) Axial velocity trends on the cut-sections: their uniformity shows that the considered geometry permits to operate the microcapillaries in parallel, as desired. d) Schematic of the porous tube fluid domain. Both the core and the spillage flow rates are indicated (only some spillages, for simplicity), together with relevant geometrical and physical entities. e) Schematic (sagittal cut-section) of the tube indicating, in particular, the axial pressure damping due to viscosity: after a characteristic length, pressure difference is dissipated so that spillage becomes negligible. We estimated such a characteristic length by introducing an analytical model, and we exploited that estimate to also set the length of the porous tube. f) Schematic of the fluid domain (relevant portion, by symmetry) used in the 3D numerical simulations carried out to assess the accuracy of the analytical model. g, h) Illustrative results showing that the analytical solution approximates the numerical one accurately enough for our purposes; both the solutions reported in Equation 5 and Equation 6 are considered, respectively in g) and h), for completeness. Explicitly accounting for relevant physical entities, the analytical solution provides a useful tool to evolve and optimize chip design.

Figure 3

Dynamic 3D real scale bio-hybrid blood-brain barrier model. a) In the left image, the 3D rendering of confocal laser scanning microscopy (CLSM) showing bEnd.3 cells (f-actin in red) cultured covering the porous tubes of the microfluidic system; the right image shows a high magnification SEM image. b) Co-culture of endothelial-like bEnd.3 and human glioblastoma U-87 cells on the porous tubes of the microfluidic system; in the left image, the CLSM of bEnd.3 and U-87 cell membranes (shown in red and in green, respectively) and, in the right image, SEM image with black arrows indicating U-87 cells. c) Immunofluorescence staining against zonula occludens-1 (ZO-1), revealing a high expression of this marker (ZO-1 in green, f-actin in red and nuclei in blue). d) Quantitative evaluation of the porous area (%) of the microcapillaries without cells and after the culture / co-culture. e) Fluorescence time-lapse images of the fluorescent dextran pumped in the microfluidic system without cells (images on top) and on the bio-hybrid system with bEnd.3 cells (images on bottom) and acquired at time t = 0, 25, 50 and 75 s. The extratubular concentration of the fluorescent dextran was monitored during time-lapse fluorescence imaging in two region of interest (ROI1 and ROI2). The extratubular dextran concentrations over time in ROI1 and ROI2 for the microcapillaries without cells and covered by bEnd.3 cells are reported in f) and g), respectively.